281 
1879.] of Physical Law. 
positive decision and a reason for it, not necessarily because 
their intellectual integrity is less, but because they cannot 
realise how vain a thing the human reason is. To imagina- 
tion and reason, controlled and checked by experiment and 
observation, are we to look as the source of the greatest 
advancement in science ; but we are not to look for infalli- 
bility, and in cases where the reason alone is allowed to 
decide, where observation and experiment are impossible, 
the uncertainty must necessarily be greater. In many cases 
the faCt that the subject is so intrinsically difficult that no 
experimental check is possible, appears to inspire the inves- 
tigator with a confidence in his conclusions that could hardly 
be reinforced by absolute certainty. 
I have confined myself to the errors which scientific men 
have committed, and to which they are liable in their search 
for truth, not because they alone are liable to err, but 
because a discussion of the multitude of errors into which 
intellectual men of other professions have fallen would be 
sure to give offence. But it is not the scientific mind which 
stands impeached — it is the human mind ! 
To what end have we then come ? It appears that all 
scientific results are attended with some uncertainty. Some- 
times the uncertainty is very small, and we are able to 
obtain a numerical estimate of it. In other cases it may be 
possible that a fundamental misconception of the truth may 
have been formed. As an instance illustrating what I mean 
we may cite the case of the measurement of the Atlanta 
base-line by the engineers of the United States Coast 
Survey. The whole length of the base-line was nearly 
6 miles, and three determinations of its length showed 
differences of about 3-ioths of an inch — about a millionth 
of the entire length. It is safe to say that if these re- 
measurements had shown differences of 10 feet there would 
have been no quarrelling in regard to which measurement 
was right, but all would have been rejected ; and if the 
engineers were not dismissed as incompetent, they would, 
with feelings of mortification, have begun their work over 
again. It is curious to observe that in many cases where 
less skilful men attack problems infinitely more complex, 
reaching conclusions differing as widely as the poles, we 
have, instead of conscientious re-investigation or a modest 
reservation of judgment, dogmatic discussions, empty words. 
In the other case, where the error is likely to be a funda- 
mental one, the probability of the truth or error of a con- 
clusion cannot always be determined numerically, and will 
vary greatly in different minds. 
